Replicates by design

This page is generated from the code and must not be edited by hand.

Not available in this version. Two pieces of work are suggested as repeats of one another when their names give the same amounts and words and at least three separate kinds of evidence agree.

replicates_by_design version 1

• Kind of rule: Pairs suggested for a look (proposes). Proposes that two files are related in one of the eight ways, for a person to look at. A proposal is never a conclusion.

• What it concludes: equivalent, named replicate_of, in respect of experimental_condition

• Reasoning: defeasible. A conclusion is usually right, and is retracted when a rule it names concludes against it.

• What it reads: a group of files

Is it available in this version?

Not available in this version. It proposes pairs but names no set of files to draw them from, so it proposes nothing yet.

What would defeat it

It names no rule whose conclusion would defeat it.

Parameters

Parameter	Shipped value	May be	What it does
agreeing_kinds	["quantity", "word"]	list of text	Which components must match for the designs to be the same. A replicate repeats the condition exactly; anything that differs makes it a variant instead.
minimum_support	3	whole number, from 1 to 8	How many distinct support families must agree. The parameter a sweep would vary, and the curve it produces reports coverage against conflict rather than a single objective, because a rule that never fires has a perfect failure rate.
requires_repeat_marker	false	yes or no	Whether an explicit repeat marker must be present. False by default because most replicates are distinguished by date rather than by a marker, and true is the setting for a corpus where the convention holds.